To enhance the adaptability of the recognition model to non-standardized data, this study designed a Wide-Baseline cattle facial recognition method based on image matching, using SuperPoint and SuperGlue to construct the image matching algorithm, and made targeted improvements to improve the recognition accuracy. In the process of feature point extraction, a dynamic threshold adjustment strategy based on image quality assessment was introduced to quantitatively assess the clarity, edge density and texture characteristics of the image, and adjust the threshold value of SuperPoint accordingly to realize high-quality feature point extraction. Meanwhile, the matching threshold of SuperGlue was adjusted by evaluating the distribution entropy and spatial coverage of feature points to achieve efficient matching. In order to verify the effectiveness of the method, a comparison test was conducted with multiple image matching methods on the self-constructed cow face dataset. The results showed that: 1) The macro average precision, macro average recall and macro average F1 of the algorithm on the Narrow-Baseline dataset were 92.1%, 90.4% and 91.2%, respectively. 2) The macro average precision, macro average recall and macro average F1 of the algorithm on the Wide-Baseline dataset were 82.8%, 81.0% and 81.9%. 3) The micro F1 score of the algorithm on the public dataset was 86.4%, and all the results were significantly better than the traditional image matching algorithm. In summary, this study aims to propose a new image matching algorithm, which effectively improves the adaptability of the cattle facial recognition model to non-standardized data, and provides technical references for the practical application of cattle facial recognition methods.
深度学习的应用显著提升了识别性能。Xu等[15]评估了RetinaNet算法在不同环境下识别牛只面部的效果,结果显示该算法在精度和速度方面表现优异,平均精度达99.8%,体现了深度学习在牛只识别和动物福利改进中的潜力。吴赛赛等[16]探究了基于深度学习的姿态估计在动物行为识别中的应用,强调了关键点检测和行为分类的重要性。Chen等[17]开发了一种特征深度网络模型(Global and part network,GPN),在奶牛面部数据集上的表现优于现有方法。但深度学习的高成本和数据需求限制了其在大规模牧场中的普及。
TongB, ZhangLi, LiG P. Progress in molecular and genetic modification breeding of beef cattle in China[J]. Genetics, 2017, 39(11): 984-1015 (in Chinese)
SunY K, WangY J, HuoP J, CuiZ Q, ZhangY G. Research progress of individual identification method of dairy cattle and its application[J]. Journal of China Agricultural University, 2019, 24(12): 62-70 (in Chinese)
[5]
LuptonSara J., SmithDavid J., ScholljegerdesEric, IveyShanna, YoungWendy, GenualdiSusan, DeJagerLowri, SnyderAbigail, EstebanEmilio, and JohnJ. Johnston. Plasma and skin per- and polyfluoroalkyl substance (PFAS) levels in dairy cattle with lifetime exposures to PFAS-contaminated drinking water and feed .[J]. Journal of Agricultural and Food Chemistry, 2022, 70(50): 15945-15954
[6]
ZinT T; MisawaS; PwintM Z; ThantS; SeintP T; SumiK; YoshidaK. Cow identification system using ear tag recognition[J]. 2020(LifeTech) IEEE 2nd Global Conference on Life Sciences and Technologies, 2020: 65-66
[7]
BiglariA, TangW. A vision-based cattle recognition system using TensorFlow for livestock water intake monitoring[J]. IEEE Sensors Letters, 2022, 6: 1-4
[8]
HuF, ZhouM, QinY L, ZhaoQ, XinY P, LiG Q. Novel cattle farm management system based on IoT and RFID[C]. In: International Conference on Computer Science and Management Technology (ICCSMT),Piscataway: IEEE Computer Society, 2022: 60-63
[9]
SimanungkalitG, BarwickJ, CowleyF, DawsonB, DobosR, HegartyR. Use of an ear-tag accelerometer and a radio-frequency identification (RFID) system for monitoring the licking behaviour in grazing cattle [J]. Applied Animal Behaviour Science, 2021, 244: 105491
[10]
WilliamsL R, FoxD R, Bishop-HurleyG J, SwainD L. Use of radio frequency identification (RFID) technology to record grazing beef cattle water point use [J]. Computers and Electronics in Agriculture, 2019, 156: 193-202
[11]
BangiyevR, ChudaevM, SchaffnerD, GoldmanE. Higher concentrations of bacterial enveloped virus Phi6 can protect the virus from environmental decay[J]. Applied and Environmental Microbiology, 2021, 87
[12]
AboubakrH A, SharafeldinT A, GoyalS M. Stability of SARS‐CoV‐2 and other coronaviruses in the environment and on common touch surfaces and the influence of climatic conditions: A review[J]. Transboundary and Emerging Diseases, 2021, 68(2): 296-312
[13]
NguyenK, FookesC, RossA, SridharanS. Iris recognition with off-the-shelf CNN features: A deep learning perspective [J]. IEEE Access, 2018, 6: 18848-18855
[14]
LeeY W, KimK W, HoangT, ArsalanM, ParkK. Deep residual CNN-based ocular recognition based on rough pupil detection in the images by NIR camera sensor[J]. Sensors (Basel, Switzerland), 2019, 19
[15]
KumarS, ChaubeM K, KumarS S. Secure and sustainable framework for cattle recognition using wireless multimedia networks and machine learning techniques[J]. IEEE Transactions on Sustainable Computing, 2022, 7: 696-708
[16]
AchourB, BelkadiM, FilaliI, LaghroucheM, LahdirM. Image analysis for individual identification and feeding behaviour monitoring of dairy cows based on convolutional neural networks (CNN)[J]. Biosystems Engineering, 2020, 198: 31-49
[17]
XuB B, WangW S, GuoL F, ChenG P, WangY W, ZhangW J, LiY F. Evaluation of deep learning for automatic multi-view face detection in cattle[J]. Agriculture, 2021
WuS S, WuJ Z, ChengG D, ZhangK, XingL W, HanS Q. Research progress of animal behavior recognition based on posture estimation[J]. Journal of China Agricultural University, 2023, 28(6): 22-35 (in Chinese)
[20]
ChenX L, YangT L, MaiK Z, LiuC X, XiongJ T, KuangY J, GaoY F. Holstein cattle face re-identification unifying global and part feature deep network with attention mechanism[J]. Animals, 2022, 12:1047
[21]
MikolajczykK, SchmidC. A performance evaluation of local descriptors[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2003, 27: 1615-1630
RubleeE, RabaudV, KonoligeK, BradskiG. ORB: An efficient alternative to SIFT or SURF[C]. In: International Conference on Computer Vision. Piscataway: IEEE Computer Society.2011: 2564-2571
[24]
Mur-ArtalR, MontielJ M M, TardósJ D. ORB-SLAM: a versatile and accurate monocular SLAM system [J]. IEEE Transactions on Robotics, 2015, 31(5): 1147-1163
[25]
Mur-ArtalR, TardósJ D. ORB-SLAM2: an open-source SLAM system for monocular, stereo, and RGB-D cameras [J]. IEEE Transactions on Robotics, 2017, 33(5): 1255-1262
ZhongH M, WangW, ZhangH H. A visual tracking algorithm combining ORB features and color models[J]. Pattern Recognition and Artificial Intelligence, 2015, 28(1): 90-96 (in Chinese)
ZhengT X, HuangS, LiY F, FengM C. Key techniques for vision based 3D reconstruction: A review[J]. Acta Automatica Scinica, 2020, 46(4): 631-652 (in Chinese)
WuR H, LiS Z, ZouF M. Image retrieval based on SIFT features[J]. Computer Application Research, 2008, 25(2): 478-481 (in Chinese)
[32]
AslanM F, DurduA, SabanciK. Human action recognition with bag of visual words using different machine learning methods and hyperparameter optimization[J]. Neural Computing and Applications, 2019, 32: 8585-8597
[33]
DeToneD, MalisiewiczT, RabinovichA. Superpoint: Self-supervised interest point detection and description [C].In:Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway:IEEE Computer Society, 2018: 224-236
[34]
SarlinP E, DeToneD, MalisiewiczT, RabinovichA. Superglue: Learning feature matching with graph neural networks[C].In:Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Computer Society, 2020: 4938-4947
[35]
QiaoY L, SuD, KongH, SukkariehS, LomaxS, ClarkC. Individual cattle identification using a deep learning based framework[J]. IFAC-PapersOnLine, 2019, 52(30): 318-323
[36]
ZhaoK, JinX, JiJ T, WangJ, HaoM, ZhuX F. Individual identification of Holstein dairy cows based on detecting and matching feature points in body images[J]. Biosystems Engineering, 2019, 181: 128-139
[37]
WengZ, LiuS Q, ZhengZ Q, ZhangY, GongC L. Cattle facial matching recognition algorithm based on multi-view feature fusion[J]. Electronics, 2023, 12(1): 156